AI Music. Is It Already Too Late To Protect Future Creativity?

Updated: Jul 8

Major Labl founder Mark Knight discusses three worrying implications of AI music maker Udio on human music creators.

AI Music. Is It Already Too Late To Protect Future Creativity?

Remember when music was free?

It wasn’t that long ago that the music industry was in crisis, Napster and Limewire were killing CD sales and music was free via file sharing and BitTorrents. However, while these services delivered on convenience they failed on quality. The stage was set for Spotify, the music- industry ‘White Knight’ who bought convenience and quality with a service worth paying for.

By 2018 approximately 20,000 new tracks were being added to Spotify each day. Now fast forward to 2022 when that number had hit 100,000. AI-generated music has arrived.

Udio changes the game for AI music creation.

Last week I was introduced to udio.com, one of the most impressive AI audio creation platforms to hit the market. With Udio anyone can create music, in any genre, in seconds, no musical talent is required. While the non-musician in me was fascinated, my thoughts quickly turned to the worrying implications for real musicians.

Here are three of my biggest fears for the future of the music industry.

1. Creative inspiration is not recognised or rewarded.

Udio works because it has already been trained on a large music catalogue. Unfortunately, we don’t know which artists were used to train the music Udio generates. This makes the task of fair royalty payments impossible.

In December. 2023 The New York Times became the first media organisation to sue OpenAI and Microsoft over the unauthorised use of published work to train artificial intelligence technologies. It’s unlikely we’ll hear the decision until January 2025. This case could provide a useful precedent but in the case of Udio there are potentially millions of defendants who don’t even know their work has been scraped. While the major label lawyers will come out swinging to defend their catalogues, who will defend independent artists?

2. The death of music sync and composition.

Getting your music placed in TV, film or games is a key revenue stream for musicians. Music supervisors will be briefed to find music that matches a brief. EG ‘We’re looking for a folk song about the summer, with a 70s vibe.” they will search for music to match that brief, creating a short-list to present back to the producers. When they can’t find what they want, musicians are commissioned to compose music to fit the brief. It’s a complicated, and time-consuming process. While the prize for landing a sync can be huge, it can be frustrating for those who fall short.

Now imagine this process for a filmmaker using AI. He needs a soundtrack, but no longer requires a budget or a music supervisor to help him. He visits Udio, and creates exactly what he wants, in minutes for free. No money is paid to musicians for their work. The music sync and compositions revenue stream disappears overnight.

3. The major labels and Spotify close the distribution door on independent artists.

The major record labels are already concerned about the volume of music being added to the streaming services. They believe this so-called ‘noise’ is taking streams and income away from their artists. Earlier this year Spotify agreed to change their payout model. Now tracks must have reached at least 1,000 streams in the previous 12 months to generate royalties on the platform.

Critics argue that this is more than a tax on AI, it’s just the first step in diluting the noise of unsigned and independent music who contribute about 96% of the music added to Spotify every day. Now with AI music services like Udio making music generation even easier, how long will it be until Spotify decides to block unsigned artists altogether, or limit distribution to a smaller number of approved partners?

This all sounds bleak. So who is the next ‘White Night’?

Last week a French tech company Ircam Amplify announced the launch of a new AI music detection tool that claims to have a 98.5% accuracy rate. While this won’t stop the creation of AI music or pay musicians for their lost royalties, it could limit the distribution of AI music, which is at least a step in the right direction.

Words Mark Knight